RE: FW: DT's review of draft-ietf-tewg-mib-06.txt

["Naidu, Venkata" <Venkata.Naidu@Marconi.com>]

Kireeti,

-> > > If you give me a list of criteria for what constitutes 
-> an interface,
-> > > I can try to say how TE tunnels differ.
-> >
-> > The simplest criteria I can describe would be:
-> > 1) If an IP address can be assigned to it, then it's an interface.

  I haven't read anything about "IP interface" in any RFC. AFAIK,
  from MIB-II perspective "interface" is referred to either 
  physical or logical interface in L1/L2. The information an 
  interface has more of L1/L2 specific (ifindex, iftype etc etc).

  MIB-II says that IP information is "augmented" over the L1/L2 
  interfaces (such as IP address, mask etc etc). Every
  router developer thinks from their design perspective and
  assumes that there is an "IP interface(s)" on top of "L1/L2
  interface".

-> I can assign an IP address to a router (router ID).  Does that make
-> routers interface?
-> 
-> > 2) Else if you can configure a route in a classic IP routing table
-> > that points across it, then it's an interface.

  For end-to-end perspective, each L1/L2 is assigned at least one
  IP address (if it is not unnumbered). So that, in strong end
  systems, an IP datagram destined to a node (router/host) must
  be accepted iff the datagram's incoming interface matches 
  the IP address of that interface.

  This is not true in non-stong-end systems.

-> That seems way too general and very arbitrary.  Can you point me at
-> an RFC that defines an interface?
-> 
-> > 3) Else it's optional as to whether it's an interface.
-> 
-> Given my druthers, I would not call TE tunnels interfaces.

  TE tunnels can be called as an "L1/L2 interface" except that
  a router acts as week-end system over such interfaces.

  This is because, TE tunnels are finally mapped to a real 
  "physical interface". If a physical interface flaps, TE
  tunnels are dynamically mapped to other interfaces (quickly
  is the intension).

  Now is that fact that IP address information can be "augmented"
  on top of "TE tunnel interfaces" along with traditional
  "interfaces".

Venkata.

PS: Kireeti, do you agree ? :)